unnecessary PREG_OFFSET_CAPTURE usage (?)

[ViliusS]

Just wondering is PREG_OFFSET_CAPTURE is needed here

Shortcode/src/Parser/RegularParser.php

Line 338 in 65d8b29

$count = preg_match_all($this->lexerRegex, $text, $matches, PREG_SET_ORDER | PREG_OFFSET_CAPTURE);

I don't see offset used anywhere in the following code however during tests I found that it uses up memory pretty heavily. If I did my testing right, in one case preg_match_all() uses up 80 MB with PREG_OFFSET_CAPTURE and 23 MB without.

[thunderer]

Interesting, it looks like a leftover from the time I actually used the offset, now replaced with calculation via mb_strlen(). Would you be interested in preparing a PR for this?

[ViliusS]

I've actually tried that, but the logic in the following switch statement is a little bit over my head. What I could share is why I have investigated this, and minimal test cases which shows memory issues with that preg_match_all() statement https://1drv.ms/f/s!AgnMn-haWyFrkcN50beuEO4A0m6PQw?e=GjwcEV . Ideally I wanted to find out why these content cases uses so much memory, but other cases don't. PREG_OFFSET_CAPTURE looked like a low hanging fruit, but there could be other issues still.

[thunderer]

@ViliusS @funkjedi the PR is merged now, and I also got another idea to make it even better using PCRE MARK verb. Unfortunately this will work in PHP 7+ only, so I'll merge it after figuring out the strategy to bump the required PHP version and modernize the codebase. My plan so far is v0.8.0 for PHP 5.3+, v1.0.0 for PHP 7.1+ and v2.0.0 for PHP 8.1+, with dedicated branches for each major version - master stays for 0.x, then v1.x, and v2.x respectively. PR is here: #116